cendres at videotron.ca
Fri Jan 2 19:52:44 PST 2004
On January 2, 2004 10:12 pm, Archaic wrote:
> On Fri, Jan 02, 2004 at 09:22:45PM -0500, ashes wrote:
> > So hlfs is tracking and testing against lfs-cvs?
> Seems the only logical thing. We are using LFS as our base, and our book
> won't be ready for a while. I also believe that if we can, we should
> stick to their package selection. However, there will be some problems,
> but we should truly try LFS packages first. Things like sysklogd may be
> able to be patchd to give the functionality we are looking for. If not,
> we need to have serious discussion weighing pros and cons. I'm not saying
> we should try and supercharge a jalopy, but we needn't be too quick to
> change packages, either. One hairy item will be HJL vs FSF. If FSF keeps
> up with the changes, or patches can be backported, we should stick with
> it, even though, personally, I prefer HJL. I'm sure no one on dial-up
> wants to see us saying "Throw package X away because we use package Y".
> Patches, however, are much less bandwidth intensive.
I can agree with good reasons to keep fsf binutils. It would mean et_dyn would
be shelved untill -pie is in stable fsf binutils, which isn't a bad idea
considering the amount of other things to do; hcc would be a snap to write if
its just for propolice. Syslog daemons are only like 150kb compressed source,
wouldn't kill anyone to download that. There's no reason we should strictly
use lfs packages, if we can find good quality syslog or even telnet clients
that support encryption, they should get good consideration.
More information about the hlfs-dev